svn commit: r315222 - in head: devel/qdevelop games/warmux graphics/xaos math/qtiplot multimedia/minitube print/hplip textproc/qstardict

[Jason Helfman]

On Wed, Apr 3, 2013 at 10:36 AM, Max Brazhnikov <makc at freebsd.org> wrote:

> On Thu, 28 Mar 2013 03:07:57 +0000 Alexey Dokuchaev wrote:
> > On Wed, Mar 27, 2013 at 10:57:34PM -0400, Eitan Adler wrote:
> > > On 27 March 2013 22:53, Alexey Dokuchaev <danfe at freebsd.org> wrote:
> > > > On Mon, Mar 25, 2013 at 01:59:18PM +0000, Max Brazhnikov wrote:
> > > >> New Revision: 315222
> > > >> URL: http://svnweb.freebsd.org/changeset/ports/315222
> > > >>
> > > >> - Remove OPTIONS_DEFINE, if it contains only global options
> > > >
> > > > Can you elaborate on this one?  Given that there is no clear policy,
> > > > personally I've been always including OPTIONS_DEFINE, even if just
> for,
> > > > say, DOCS.
> > >
> > > Options such as IPV6, DOCS, EXAMPLES, and NLS are generally set
> > > globally (e.g., OPTIONS_UNSET= NLS) and not per port.  As such, it can
> > > reduce user annoyance if the dialog box is not shown for such options.
>
> and it keeps status quo with old options framework, when DOCS/EXAMPLES
> were not
> usually added to the options list, apparently because bsd.port.mk checked
> only
> for NOPORT* variables.
>
> > Yes, I've heard that reasoning.  I'm not particularly against it, but I
> > am also worried about certain inconsistency: it means, that for a port
> > with at least one "local" option, users can toggle all of them, while for
> > others they had to rely on their /etc/make.conf settings (or defaults).
>
> I agree with you about inconsistency. I'm fine with any policy once it's
> settled.
>
> Max
>
>
I'm fine with any policy, however it seems that there is inconsistency in
what I've heard for defining, as well.
Why not have a policy first, then implement/document?

-jgh

--
Jason Helfman          | FreeBSD Committer
jgh at FreeBSD.org     | http://people.freebsd.org/~jgh  | The Power to Serve